ISciences Open Science Portfolio

Advancing Earth Science Research Through Open Tools and Education

Joshua Brinks

Research Scientist, ISciences, LLC
jbrinks@isciences.com

Presentation Outline

  • ISciences Overview
  • DANTE: Data Analysis Toolbox for Environmental Research
  • TOPS-SCHOOL: NASA’s Open Science Training Curriculum
  • SCALAR: AI Literacy for Earth Scientists
  • nClimGrid Explorer: Interactive Climate Data Visualization
  • exactextractr: High-Performance Zonal Statistics Package
  • GDAL Contributions: Enhancing Open Geospatial Tools
  • WSIM: Water Security Indicator Model

ISciences

Who We Are

  • FOSS Focused Research and technology firm based in Burlington, VT
  • Specializing in geospatial analytics and applied statistical modeling
  • Focus on water resources, political instability environmental monitoring, and risk assessment
  • Small team of scientists, engineers, and developers

What We Do

  • Partner with agencies including NASA, NOAA, USACE, intelligence
  • Develop water security monitoring systems (WSIM)
  • Develop open source software, packages, and educational materials
  • Support decision-making through data-driven insights

Broader Benefits

  • Roadmaps for access to data and tools
  • Reduce barriers for learning these tool
  • Enable participation from diverse stakeholders
  • Solicit feeback from under-resourced communities
  • Builds trust through transparency

DANTE

Data Analysis Toolbox for Environmental Research

Overview

  • Collaborative open-science platform developed by USACE, ISciences, Columbia’s CIESIN, and CASE International
  • Accelerates multidisciplinary environment-security applied research
  • Bridges political science, social science, geography, environmental sciences, and defense communities

Access

TOPS-SCHOOL

ScienceCore Heuristics for Open Science Outcomes in Learning

Overview
  • NASA-funded Open Science training curriculum
  • Joint effort between CIESIN, ISciences, CUNY
  • Emphasizes hands-on learning with real NASA datasets
  • Cloud-based learning environment/workshops (2i2c)
  • Case studies using recent environmental events
  • 4 Modules: Water Module, Air Quality, Disasters, Climate and Agriculture

TOPS-SCHOOL Case Studies

SCALAR

ScienceCore Curriculum for AI Literacy and Research

Overview

  • Proposed NASA HPOSS-funded follow-on to TOPS-SCHOOL
  • Addresses AI literacy challenges in Earth Sciences
  • Teaches responsible integration of LLMs with NASA data
  • Focuses on recent missions: TEMPO, ECOSTRESS, SPORT-LIS, SWOT, PACE
  • Status: Seeking funding and institutional home

Learning Outcomes

  • Access NASA Earth Science datasets with AI assistance
  • Design effective prompts for scientific contexts
  • Recognize and mitigate AI hallucinations and bias
  • Validate AI-generated code and analyses
  • Apply open science principles to AI workflows

nClimGrid Explorer

Interactive Climate Data Visualization Tool

  • R package and Shiny web application for NOAA nClimGrid-Daily dataset
  • Funded by ESIP Lab grant program
  • Provides open-science tools for climate and social vulnerability analysis

nClimGrid Explorer

Applications

  • Climate researchers and emergency managers
  • Policy makers and practitioners
  • Drought monitoring and risk assessment
  • Environmental justice analyses

exactextractr Package

High-Performance Zonal Statistics for R

Overview

  • Fast and accurate zonal statistics extraction from raster data
  • Orders of magnitude faster than competing methods (raster::extract, velox, ESRI, QGIS, GRASS, rasterio)
  • Implements precise algorithms for area-weighted statistics
  • Coverage fraction calculations for accurate area-weighted statistics

Main Website | R Package Website | Python Module Website

GDAL Contributions

Enhancing Open Geospatial Tools

About GDAL

  • Industry-standard geospatial data abstraction library
  • Powers countless GIS applications and tools worldwide
  • Supports 200+ raster and vector formats
  • Used by QGIS, ArcGIS, Google Earth Engine, and more
  • Open-source (MIT/X license)

Latest Contribution: gdal raster zonal-stats

  • New command in GDAL 3.12.0 “Chicoutimi” (upcoming release)
  • Brings exactextractr-style performance to native GDAL
  • Fast, accurate area-weighted zonal statistics
  • C++ implementation integrated into core GDAL
  • Part of the modern gdal unified command interface

Key Features

  • Three pixel methods: default, all-touched, fractional coverage
  • 20+ statistics: mean, sum, min/max, variance, weighted variants
  • Weighting support: population-weighted calculations
  • Pipeline integration: combine with other GDAL operations
  • Works with polygons or categorical rasters
  • Handles different resolutions automatically

Impact

  • Fills critical gap in GDAL utilities
  • No Python dependencies required
  • Command-line accessible for all users
  • Docs

WSIM: Water Security Indicator Model

Overview

  • Global water security monitoring and forecasting system
  • Operating continuously since April 2011
  • Monthly reports with 1-9 month lead-times
  • Validated against subsequently observed data
  • Regional blogs detail monthly conditions globally

Technical Approach

  • Data Source: ECMWF Reanalysis v5 (ERA5)
  • Resolution: Quarter-degree (upgraded from half-degree)
  • Baseline Period: 1981-2020 (satellite era focus)
  • Forecast Model: NOAA CFSv2 ensemble forecasts
  • Composite anomaly index based on soil moisture, ET deficit, runoff, and blue water

Key Features

  • Return period methodology (3-40+ year events)
  • Five severity categories: abnormal to exceptional
  • Depicts both deficits (red) and surpluses (blue)
  • Purple shading indicates concurrent deficit/surplus

Current Status & Future

  • Monthly regional blogs available
  • Detailed reports available via ISciences
  • Seeking: NASA INNOVATE grant funding
  • Vision: Open cloud-based version for public access
  • Contact for partnership opportunities

Open Science Impact

Benefits

  • Roadmaps for data and tool access
  • Reduced barriers to learning
  • Diverse stakeholder participation
  • Community feedback integration
  • Transparency builds trust

Principles

  • Open data and code
  • Reproducible workflows
  • Community-driven development
  • Comprehensive documentation
  • Educational resource sharing

Thank You

Acknowledgments

  • NASA TOPS/Open Science
  • Center for International Earth Science Information Network (CIESIN) - Columbia University
  • Open Science Team and SCHOOL contributors
  • Open source community contributors

All materials available under open licenses. Visit our repositories for code, data, and documentation.